11.3 Bernoulli-Logit Generalised Linear Model (Logistic Regression)
Stan also supplies a single primitive for a Generalised Linear Model with Bernoulli likelihood and logit link function, i.e. a primitive for a logistic regression. This should provide a more efficient implementation of logistic regression than a manually written regression in terms of a Bernoulli likelihood and matrix multiplication.
11.3.1 Probability Mass Function
If \(x\in \mathbb{R}^{n\cdot m}, \alpha \in \mathbb{R}^n, \beta\in \mathbb{R}^m\), then for \(y \in {\{0,1\}}^n\), \[\begin{align*} &\text{BernoulliLogitGLM}(y~|~x, \alpha, \beta) = \prod_{1\leq i \leq n}\text{Bernoulli}(y_i~|~\text{logit}^{-1}(\alpha_i + x_i\cdot \beta))\\ &= \prod_{1\leq i \leq n} \left\{ \begin{array}{ll} \text{logit}^{-1}(\alpha_i + \sum_{1\leq j\leq m}x_{ij}\cdot \beta_j) & \text{if } y_i = 1, \text{ and} \\ 1 - \text{logit}^{-1}(\alpha_i + \sum_{1\leq j\leq m}x_{ij}\cdot \beta_j) & \text{if } y_i = 0. \end{array} \right. \end{align*}\]
11.3.2 Sampling Statement
y ~ bernoulli_logit_glm(x, alpha, beta)
Increment target log probability density with bernoulli_logit_glm_lpmf( y | x, alpha, beta)
dropping constant additive terms.
11.3.3 Stan Functions
real bernoulli_logit_glm_lpmf(int[] y | matrix x, real alpha, vector beta)
The log Bernoulli probability mass of y given chance of success
inv_logit(alpha+x*beta), where a constant intercept alpha is used
for all observations. The number of rows of the independent variable
matrix x needs to match the length of the dependent variable vector
y and the number of columns of x needs to match the length of the
weight vector beta.
real bernoulli_logit_glm_lpmf(int[] y | matrix x, vector alpha, vector beta)
The log Bernoulli probability mass of y given chance of success
inv_logit(alpha+x*beta), where an intercept alpha is used that is
allowed to vary with the different observations. The number of rows of
the independent variable matrix x needs to match the length of the
dependent variable vector y and alpha and the number of columns of
x needs to match the length of the weight vector beta.